Skip to content

Evaluation Script and ORE Corpus Creation #20

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
martasoricetti wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Evaluation Script and ORE Corpus Creation #20

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f762487
- Evaluation Code
martasoricetti Jan 14, 2026
2873e6d
ORE readme.md
martasoricetti Jan 14, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
106 changes: 67 additions & 39 deletions extractor/cex/combined.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -230,6 +230,70 @@ def find_sentences_in_figure(self, figure, dict_to_check, ns, formula_map, super

return []

def return_heads_to_use(self, head_elements):
roman_pattern = re.compile('''^((?=.)M{0,4}(CM|CD|D?C{0,3})(XC|XL|L?X{0,3})(IX|IV|V?I{0,3})\.*)([A-Za-z0-9]+\.*)*\s+''')

# Global stats
head_texts = [(h.text or "").strip() for h in head_elements]

head_n_attribute = sum("n" in h.attrib for h in head_elements)
head_no_n_attribute = sum("n" not in h.attrib for h in head_elements)

head_roman = sum(1 for t in head_texts if roman_pattern.search(t))
head_no_roman = sum(1 for t in head_texts if not roman_pattern.search(t))

use_n_attribute = head_n_attribute >= (head_no_n_attribute + head_n_attribute) / 2
use_roman_numeration = head_roman >= (head_no_roman + head_roman) / 2

heads_to_use = dict()

for i, head in enumerate(head_elements):
head_text = (head.text or "").strip()

# 1) Case: numeric @n is the dominant convention
if use_n_attribute:
if "n" in head.attrib:
n = head.get("n", "")
n_parts = [p for p in n.split(".") if p]

# First-level only: e.g. "1", "2" but not "1.1"
if len(n_parts) == 1:
# Remove leading numeric pattern from the visible heading text
if re.search(r'\b\d+(\.\d+)+\b', head_text):
split_string = re.split(r'\b\d+(\.\d+)+\b', head_text)
split_string = [s.strip() for s in split_string if s.strip()]
if split_string:
heads_to_use[head_text] = split_string[0]
else:
heads_to_use[head_text] = head_text
else:
heads_to_use[head_text] = head_text
else:
# Heuristic: skip “lonely” heads if neighbors have @n
prev_has_n = i > 0 and "n" in head_elements[i - 1].attrib
next_has_n = i < len(head_elements) - 1 and "n" in head_elements[i + 1].attrib
if prev_has_n and next_has_n:
continue
# keep some special cases (e.g. Availability)
else:
heads_to_use[head_text] = head_text

# 2) Case: Roman numerals are the dominant convention
elif use_roman_numeration:
m = roman_pattern.search(head_text)
if m:
prefix = m.group().strip()
cleaned = head_text.replace(prefix, "", 1).strip()
heads_to_use[head_text] = cleaned or head_text
else:
heads_to_use[head_text] = head_text

# 3) Fallback: just use the text as is
else:
heads_to_use[head_text] = head_text

return heads_to_use

def convert_to_json(self):

tree = etree.parse(self.xml_file)
Expand All @@ -243,17 +307,7 @@ def convert_to_json(self):
# Precompute common elements
ref_citkey_dict = self.build_dict_ref_citkey(tree, ns)
head_elements = root.findall(".//tei:div/tei:head", namespaces=ns)
has_n_attribute = any("n" in head.attrib for head in head_elements)
roman_pattern = r'^M{0,3}(CM|CD|D?C{0,3})(XC|XL|L?X{0,3})(IX|IV|V?I{0,3})\.*\s+'
has_roman_numeration = any(re.search(roman_pattern, (head.text or "").strip()) for head in head_elements)
head_n_attribute = sum("n" in head.attrib for head in head_elements)
head_no_n_attribute = sum("n" not in head.attrib for head in head_elements)
head_roman = sum(
1 for head in head_elements if re.search(roman_pattern, (head.text or "").strip())
)
head_no_roman = sum(
1 for head in head_elements if not re.search(roman_pattern, (head.text or "").strip())
)
heads_to_use = self.return_heads_to_use(head_elements)

# Find notes with in-text reference pointers
notes_text_with_refs = set()
Expand All @@ -278,34 +332,8 @@ def convert_to_json(self):
head = div.find("./tei:head", namespaces=ns)
if head is not None:
head_text = head.text.strip() if head.text else ""
if has_n_attribute:
if head_n_attribute >= (head_no_n_attribute+head_n_attribute)/2:
if 'n' in head.attrib:
n = head.get("n")
n_parts = n.split(".")
if '' in n_parts:
n_parts.remove('')
if len(n_parts) == 1:
if re.search(r'\b\d+(\.\d+)+\b', head_text):
split_string = re.split(r'\b\d+(\.\d+)+\b', head_text)
split_string = [substring.strip() for substring in split_string]
last_head = split_string[0].strip()
else:
last_head = head_text
else:
last_head=head_text
elif has_roman_numeration:
if head_roman >= (head_no_roman+head_roman)/2:
if re.search(r'^M{0,3}(CM|CD|D?C{0,3})(XC|XL|L?X{0,3})(IX|IV|V?I{0,3})\.*\s+', head_text):
pattern = re.compile(r'^M{0,3}(CM|CD|D?C{0,3})(XC|XL|L?X{0,3})(IX|IV|V?I{0,3})\.*\s+')
match = pattern.search(head_text)
if match:
match = match.group().strip()
last_head = head_text.replace(match, "")
else:
last_head=head_text
else:
last_head = head_text
if head_text in heads_to_use:
last_head = heads_to_use.get(head_text)
else:
head_text= f'''Section Untitled {i}'''
last_head = head_text
Expand Down
Loading